- Dillo-dev - lists.mailman3.com

[patch] send_dpip_tag in vsource.c is unused

by p37sitdu＠lavabit.com

Attached patch removes unused function send_dpip_tag. Functions send_plain_text, send_numbered_text and send_html_text are made static.

Dec. 20, 2012

Dpi_connect_socket retry

by p37sitdu＠lavabit.com

On Sun, Dec 16, 2012 at 07:43:19PM -0300, Jorge Arellano Cid wrote: > On Sun, Dec 09, 2012 at 02:58:43PM +0400, p37sitdu at lavabit.com wrote: > > Also, comment says "the server closes sock_fd on error". However, > > sock_fd should still be closed by client to free allocated file > > descriptor. OS don't know if Dillo is going to read from it (in which > > case error such as ECONNRESET should happen) so file descriptor can't > > be reused. > > > > File descriptor should be closed in case of error in a_Dpip_build_cmd > > too. > > The file descriptor is properly closed by its CCC afterwards > (Visible by defining VERBOSE in chain.c). > > The CCC manages the error propagation/handling, in both the > requesting and answer branches. File descriptor is properly closed if there were no errors. But if there is an error in a_Dpip_build_cmd or Dpi_blocking_write, sock_fd should be closed before returning from Dpi_connect_socket. Dpi_connect_socket calls a_Dpip_build_cmd to build command and Dpi_blocking_write to write this command to sock_fd. a_Dpip_build_cmd does not accept sock_fd as its argument so it can't close it. Dpi_blocking_write only writes to fd, but does not close it on error. So if one of these two function fails, Dpi_connect_socket should close sock_fd before returning -1. In the current state it just returns -1 and leave sock_fd open. By the time Dpi_connect_socket returns, sock_fd is lost and will not be closed by any code outside Dpi_connect_socket. Attached patch adds code to close sock_fd in case of error.

Dec. 20, 2012

Patch for xembed (updated for resize)

by podarcis＠lavabit.com

Hi, well the aim would be to contribute to dillo to make it embeddable by X11 applications. If you want to achieve this with "standard" XEMBED protocol only, I am off -that's not my domain... I have tested/debugged the current code with 'tabbed' (as you suggested) and found out, that it works even better if the ClientMessage code (and this is the only relevant code here) is removed. tabbed also showed some glitches with the current code: - dillo's File menu opens at different positions after moving - clicking in the grey space right to dillo's 'Tools' icon seems to crash something... So I removed the whole thing. Did you try my patch? It certainly works not only from FLTK applications. You can verify that, if you try to embed dillo into any window on your desktop e.g. a gnome terminal (use xwininfo to find out its xid). I do not claim that my patch is complete already, but it may at least be used as a starting point. At least it has the advantage, that the embedder does not need to do anything, except passing an xid. If keeping the old code is your only concern, then I can re-integrate it into my patch again - it has no effect on it at all. Cheers chris > Hi, > > On Sat, Dec 15, 2012 at 12:04:35PM +0100, podarcis wrote: >> Hi, >> >> here is an updated version making resize work. >> It still works with tabbed too. >> Note that I have had no knowledge of X or xembed at all, so all >> changes come from experimenting... >> I removed large parts of the existing code, because they did nothing >> in my environment. >> >> Please have a look at it. > > hm, the patch basically removes support for the xembed protocol. Is > it meant as a diagnostic patch to investigate a problem, or meant > for inclusion into the dillo repository? > It won't work for the latter case. You can't just remove code that > is not needed for your specific case. > If you just want to get embedding of dillo into an fltk based window > working, you would need to send the proper xembed X11 event from your > embedding fltk program. Look at what tabbed is doing. > I can also dig into my old stuff, I think I once had a minimal fltk > based embedder working. > > Cheers, > Johannes > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.auriga.wearlab.de/cgi-bin/mailman/listinfo/dillo-dev >

Dec. 20, 2012

Extensions for hyphenation

by corvid＠lavabit.com

The text spills over into the following cell in <table border=1> <tr> <td>W d<whatever>aaaaaaaaaaaa*<br> <td>WW d<whatever>aaaaaaaaaaaa*<br> <td>WWW d<whatever>aaaaaaaaaaaa*<br> <td>WWWW d<whatever>aaaaaaaaaaaa*<br> <td>WWWWW d<whatever>aaaaaaaaaaaa*<br> <td>WWWWWW d<whatever>aaaaaaaaaaaa*<br> </table> if I make the browser window narrowish.

Dec. 20, 2012

config.h.in and configure.in

by jcid＠dillo.org

On Wed, Dec 19, 2012 at 01:24:21PM +0000, Jeremy Henty wrote: > > I thought that we agreed[1] to remove config.h.in from the repository > and to rename configure.in to configure.ac . Can I make these changes > or is more discussion needed? Go ahead! -- Cheers Jorge.-

Dec. 19, 2012

config.h.in and configure.in

by onepoint＠starurchin.org

I thought that we agreed[1] to remove config.h.in from the repository and to rename configure.in to configure.ac . Can I make these changes or is more discussion needed? Regards, Jeremy Henty [1]: http://lists.auriga.wearlab.de/pipermail/dillo-dev/2012-December/009618.html

Dec. 19, 2012

Patch for xembed (updated for resize)

by johannes.hofmann＠gmx.de

Hi, On Sat, Dec 15, 2012 at 12:04:35PM +0100, podarcis wrote: > Hi, > > here is an updated version making resize work. > It still works with tabbed too. > Note that I have had no knowledge of X or xembed at all, so all > changes come from experimenting... > I removed large parts of the existing code, because they did nothing > in my environment. > > Please have a look at it. hm, the patch basically removes support for the xembed protocol. Is it meant as a diagnostic patch to investigate a problem, or meant for inclusion into the dillo repository? It won't work for the latter case. You can't just remove code that is not needed for your specific case. If you just want to get embedding of dillo into an fltk based window working, you would need to send the proper xembed X11 event from your embedding fltk program. Look at what tabbed is doing. I can also dig into my old stuff, I think I once had a minimal fltk based embedder working. Cheers, Johannes

Dec. 19, 2012

dpid patches

by p37sitdu＠lavabit.com

First patch is for removing expired TODO item, just resending it. Second patch is new, for dpid.c. Function fill_services_list has a bug: on line 463 there is a check *services_list != NULL. If it is true, then -1 is returned, but dpidrc_stream is not closed. Error message on line 456 says "popen failed", should be "fopen failed". fill_services_list checks for availability of user_dpidir and sys_dpidir, but after line 449 these variables are not used. They are just dFreed in the end. Looks like code was just copypasted from register_all. fill_services_list is only called from main.c:248 and dpid.c:807 of dpid.c. register_all is called from main.c:235 and dpid.c:806. So fill_services_list is always called after register_all and doing the same check is useless. On line 503 there is a conversion of services_alpha_comp to type (dCompareFunc). I changed the function services_alpha_comp to the usual style of writing comparison functions so no conversion is required. In the current state it is undefined behaviour. [1] [1] http://stackoverflow.com/questions/559581/casting-a-function-pointer-to-ano…

Dec. 18, 2012

[patch] Gopher DPI

by p37sitdu＠lavabit.com

Fixed memory leak related to dGetline.

Dec. 17, 2012

[patch] Gopher DPI

by p37sitdu＠lavabit.com

Small changes - Added 'p' filetype for images (used by some gopher servers) - Use dGetline instead of getline

Dec. 17, 2012